Loading
Life sciences · Preprint
arXiv · September 4, 2026
Posted before peer review. The findings may change or fail to hold.
RegionFed is a proposed federated learning framework designed to personalize query understanding across heterogeneous retail regions while preserving privacy. The method reportedly achieves 92.27% accuracy on Amazon ESCI dataset, approaching a centralized upper bound (92.04%) while providing differential privacy (ε≈0.60). This is a preprint computer science contribution not yet peer reviewed and not applicable to clinical settings.
Preprint. Retail search queries from heterogeneous geographic regions; benchmarked on public e-commerce datasets.. Intervention: RegionFed: federated learning framework operating at gradient level with adaptive personalization routing based on regional-global gradient conflict.. Compared with: Centralized model with regional weighting; standard federated learning; parameter-level personalized FL methods..
RegionFed-Meta achieves 92.27% accuracy on Amazon ESCI, with a gap of 0.23 percentage points to centralized upper bound (92.04%) Parameter-level personalization methods collapse on T5 transformers to below 10% accuracy due to tied embeddings and LayerNorm interactions Method provides (ε≈0.60)-differential privacy and O(1/√T) convergence rate
Safety was not reported in the material analysed. Check the source before drawing any conclusion about harm.
The source did not state who this applies to in practice.
This is an arXiv preprint describing a novel federated learning method for retail search systems; it has not undergone peer review and reports computational/algorithmic results rather than clinical outcomes.
As stated by the source record.
Quoted from the source exactly as published.
Graded across the dimensions that decide whether you should act, each from what the source actually supports. There is no single score, and where a dimension was not assessed it says so.
Retail search systems serve diverse geographic regions with distinct query patterns, vocabularies, and product preferences, creating significant data heterogeneity that challenges both privacy-preserving training and model personalization. Federated learning offers a natural solution for privacy, but standard FL methods produce global models that sacrifice regional performance, while existing personalized FL approaches operate at the parameter level and catastrophically collapse on modern transformers (below 10\% accuracy on T5) due to tied embeddings and LayerNorm interactions. We introduce RegionFed, an \textit{architecture-robust} federated learning framework that sidesteps this failure by operating entirely at the gradient level. RegionFed uses the $\ell_2$ conflict between regional and global gradients as a unified signal that (i) diagnoses heterogeneity, (ii) routes each region to the cheapest sufficient personalization strategy, and (iii) adaptively controls personalization strength. Because it treats models as differentiable black boxes, RegionFed deploys on T5-Small, T5-3B, RoBERTa, and CNN with zero code changes, providing large gains on transformers (where parameter-level methods collapse) and consistent improvements on CNNs. Across three public datasets (Amazon ESCI, Amazon Reviews, LEAF-FEMNIST) and four architectures, RegionFed-Meta achieves 92.27\%, closing the gap to the privacy-violating centralized upper bound (Centralized + Regional Weighting: 92.04\%, $Δ$=0.23pp, within 1$σ$) while providing $(ε{\approx}0.60)$-differential privacy and $\mathcal{O}(1/\sqrt{T})$ convergence.
Taken from the source record, never inferred. Follow any of these and new work involving them reaches your briefing.